The Emergence and Growth of Evolutionary Art –
1980–1993
Nicholas Lambert, William Latham, Frederic Fol Leymarie
Abstract
One of the most interesting—if frustrating—aspects of charting the history of computer art is trying to
understand the intersections of specific technologies and artistic experimentation. It is rarely as clearcut as a simple linear influence of one to the other, partly because artists are able to envision all kinds
of possibilities that technology might enable them to realize in some kind of form, but as they do so, the
technology is itself shaped, especially in terms of how it is perceived by others. Do artists find a way to
give technologies an aesthetic outlet, or do some technologies possess—or facilitate—a characteristic
aesthetic that finds its expression through specific artists? Certainly, in the history of computer art
it would seem that particular aesthetics, technologies, and artists are closely intertwined in certain
Nicholas Lambert
Department of History of Art,
Film and Visual Media
Birkbeck College
University of London
United Kingdom
n.lambert@bbk.ac.uk
William Latham
Goldsmiths College
University of London
United Kingdom
w.latham@gold.ac.uk
Frederic Fol Leymarie
Goldsmiths College
University Of London
United Kingdom
f.leymarie@gold.ac.uk
periods. This intertwining of art, technology, and ideas stolen from the natural world has never been so
arguably merged as the period in the history of computer art from 1980 to 1993. We take as the defining
start of this period the initial work of Mandelbrot on fractals that became known as the Mandelbrot set
and led to his famous illustrated art-science book The Fractal Geometry of Nature. In 1993, this first
highly creative period in evolutionary computer art came to an end with major publications by pioneers
Karl Sims, Stephen Todd, and William Latham.
Artist-Researchers and New Graphics Technologies in the 1980s
Although the emergence of personal computers by 1980 enabled a new generation of artists to
start experimenting with digital images, especially due to the Apple Mac, Commodore Amiga,
and Atari ST, in 1984–85, as advanced graphics and sound capabilities became available, there
were those who continued to be closely associated with academic and corporate research centres.
The Xerox PARC model was undoubtedly influential in providing a template for research
collaborations, and IBM in particular supported some interesting developments. In such
environments, artists could leverage the power of multiple networked computers, use software
that was still under development, and utilize video displays and printers that were far more
advanced than those available elsewhere.
The emergence of computer graphics as a major aspect of commercial TV and film production,
in addition to its use for scientific visualizations and military simulators, drove the development
of new graphics technologies. One key area was the simulation of natural landforms, vegetation,
seascapes, and other environmental features. Hitherto, the public perception of computer
graphics (as represented in films such as “Tron” and “War Games”) had been of vector 3D shapes
and textured solid models. However, during the 1980s, a new and increasingly ubiquitous image
appeared: the “fractal,” a word coined by Benoit Mandelbrot and visualized as the Mandelbrot
set [1, 2]. This radically changed the idea that computer graphics had to look artificial; along with
other techniques, it raised the possibility of simulating the natural world. Mandelbrot himself
referred to the collision of abstraction and naturalism that occurred when fractals were first used
by Richard Voss and others to generate landscapes:
Leonardo, Vol. 46, No. 4, pp. 367–375, 2013 | © 2013 Nicholas Lambert, William Latham, Frederic Fol Leymarie
367
What a profound irony that this new geometry, which everyone seems spontaneously
to describe as “baroque” and “organic”, should owe its birth to an unexpected but
profound new match between those two symbols of the inhuman, the dry, and the
technical: namely, between mathematics and the computer [3].
Such a possibility had actually been raised back in the early 1950s by Alan Turing himself [4].
Fascinated as he was by the symmetries and structures produced by organisms, he realized he
could investigate the area of morphogenesis using the new Ferranti computer. Unfortunately,
his death occurred in the middle of this ground-breaking work, and his thoughts on
evolutionary systems cannot be deduced from his surviving notes. Yet he pre-empted a whole
area of biological and genetic research in the 1980s that was also preceded by John Conway’s
Game of Life [5].
It was The Blind Watchmaker by biologist Richard Dawkins that inspired the first true
flourishing of artificial life (A-Life) in the computational medium, and crucially gave rise to a
number of aesthetic innovations [6]. Dawkins named the new emergent forms after the surrealist
paintings made by Desmond Morris that contained “vaguely animal-like shapes,” which Morris
described as “biomorphs” [7]. Dawkins created genetic rules for each tree that enabled them to
mutate as different lineages were bred together. He made the tree-like forms symmetrical about
their vertical axis for reasons of parsimony and aesthetics, as well as “because [he] was hoping to
evolve animal-like shapes, and most animal bodies are pretty symmetrical” [8].
Dawkins noted that despite experimenting with other growth patterns, the symmetrical plan
generated the most interesting, and indeed “lifelike,” results. This was a lucky outcome, but later
Dawkins deliberately added genes to control the segmentation of the forms, mirroring the
importance of segmented bodies in the animal kingdom [9]. With segmentation and symmetry
combined, the resulting forms had the greatest diversity and therefore “fitness.” Yet the original
decision was fortuitous, and the outcomes truly emergent.
It would seem that both Dawkins and Mandelbrot, though scientists by training and vocation,
were well aware of the aesthetic potentials of their discoveries. In Mandelbrot’s case, the
similarities with natural forms were striking and obvious, especially after they were applied to
textures and landscapes in 3D programs by Alan Norton and others [10]. However, Dawkins’
evolutionary forms would encounter the work of an artist already immersed in the concepts of
mutation and generation.
The Early Computer Artworks of William Latham
William Latham started working with computers in 1985 after completing his MA degree in fine
art at The Royal College of Art in London. At an early stage in his career, he took the concept of
evolving forms and freely developed it into a distinctive artistic style, which incorporates natural
and artificial elements. Latham is also interesting for having gradually moved away from art in
1993 and into computer games, which incorporate ideas and code taken from his earlier art work.
Latham was interested in the evolution of form even before he discovered computing. "During
the period 1983 to 1985, using a set of rules, he designed and termed “FormSynth,” for the
transformation of shapes, he set out to sketch huge drawings of multiplying, changing forms
(Figure 1).
The logic and consistency of Latham’s possible worlds arises from his concept of an evolutionary
approach to the making of sculpture. The complexity and vitality of the forms he devised is
368
Lambert, Latham, Fol Leymarie | The Emergence and Growth of Evolutionary Art – 1980-1993
Figure 1. FormSynth evolutionary drawing, William Latham. 1983-85. Details of a two-meter, hand-drawn FormSynth tree. ©
1983-85 William Latham.
Figure 2. FormGrow/Mutator Generated Art, 1989. Nine mutations (ribbed branched structures),
evolved forms resulting from the use of a Mutator session within FormGrow. © 1989 William Latham.
derived from the step-by-step accretion of “operations” on simple initial forms such as cones,
spheres, or tori [11].
John Lansdown, in his introduction to “The Conquest of Form”, Latham’s 1989 exhibition at the
Arnolfini in Bristol, UK, considered that Latham and others working with him (in particular
Stephen Todd and Mike King) were exponents of “another form of sculpture” [12]. This is
derived from the illusory yet real appearance of these works: their seeming materiality is defeated
by the obvious departures from our physical reality (Figure 2). Even this system could produce
unexpected results [13].
The Emergence and Growth of Evolutionary Art – 1980-1993 | Lambert, Latham, Fol Leymarie
369
In 1987, Latham was appointed artist and research Fellow at IBM’s UK Research Labs in
Winchester, where he began working with mathematician and programmer Stephen Todd on a
system called FormGrow. This built on FormSynth and allowed simple construction rules such
as bulging and hollowing objects. As he worked with this system, building up a new library of
rules, Latham realized that some of his long repeated sequences of FormSynth operations could
be condensed into new rules, such as those for growing tendrils and horns.
Later on, by the end of 1988, Latham and Todd started using an “add on” new system they
developed, called Mutator, that managed the data from FormGrow and began to cross-breed
forms together, by identifying their basic components as “genes” and allowing these to be
recombined and modified to produce large evolutionary trees of computer-generated imaginative
3D forms (Figure 2). As Latham says: “Mutator derives its methods from processes of nature,
and was partly inspired by a simulation of natural selection” [14]. Importantly, the Mutator
system enabled the artist to pick, breed, and marry natural-looking forms at will based on their
aesthetic quality, which gave the artist a highly intuitive and minimal interface.
This system has an overall appearance that could be called “organic” and seemingly aims toward
natural yet fantastical forms. Latham’s stylistic decision on which operations to use was made at
the level of the core design of the program itself, ensuring that all images bear his imprint, to a
degree. The aesthetic of these images, while inspired by nature and science-fiction, remains very
much their own. These are forms that would have been inconceivable without the computer to
perform all the millions of possible changes, transformations and developments that Latham
foresaw (Figure 3).
Figure 3. Mutation X. Raytraced. 1989. Final evolved form resulting from a Mutator session. © 1989 William
Latham.
370
Lambert, Latham, Fol Leymarie | The Emergence and Growth of Evolutionary Art – 1980-1993
One might justifiably question the artist’s role in images that are not merely assembled by the
computer in its capacity as a tool, but generated directly by it. Where is the human input?
Latham counters that his input is at the level of the software itself, which imparts not only the
formal construction of the objects but also more intangible aspects of their style as well.
Latham’s major influences are science-fiction imagery—one is reminded especially of H.R.
Giger’s designs for the “Alien” films [15] and the seminal book On Growth and Form by biologist
D’Arcy Thompson, which revealed the mathematical structure underpinning the shapes of life
forms [16]. By shaping the code that in turn creates the images, Latham’s involvement is at the
conceptual rather than practically artistic level, though he also exercises further artistic judgment
in choosing certain paths for his images to grow and develop, and their final visual quality in
terms of color and texture.
Karl Sims and Artificial Life
Although Richard Dawkins had given impetus to simulations of evolution, it was not until
Dawkins himself introduced his “Biomorphs” that a number of major figures in this emerging
area, including Karl Sims, began developing their own concepts [17].
From this, Sims developed a program that used the basic genetic concepts of selection,
reproduction, and sexual combination to evolve an artificial genotype, and in so doing represent
the Darwinian idea of “fitness” in action (Figure 4).
Figure 4. Extinct Image, Karl Sims, 1990. From “Artificial Evolution for Computer Graphics,” ACM SIGGRAPH '91 Conference
Proceedings, Las Vegas, Nevada, July 1991 (c) Karl Sims.
“A-Life, then, is fundamentally concerned with understanding and formalizing the underlying
dynamic structures of living things” [18]. A-Life art, defined in the simple sense outlined above,
engages with the same ideas: it is a form of art practice that begins to take on, in various ways,
the abstract dynamics of nature.
The Emergence and Growth of Evolutionary Art – 1980-1993 | Lambert, Latham, Fol Leymarie
371
In terms of the creative process, both Sims and Latham primarily use the human as a cost
function in a Monte Carlo-type simulation to explore a multi-dimensional parameter space in
which the unevenness of parameter space is covered. Interestingly, during this period, the use of
constraints played a small role, apart from individual mutate-able parameters with maximum
and minimum values and the implicit constraints of Sims’ Lisp Code and Latham’s FormGrow
logic. In addition, the automatic culling of generated forms (based on mathematical criteria) to
reduce the size of the search space was limited. Replacing the artist (as a fitness function selector)
was also very limited.
In his 2001 essay that considers the origins of A-Life and art, Mitchell Whitelaw points to a 1987
seminar by Christopher Langton in Los Angeles as the starting point for this area. Looking for
precedents in art for the concept of artificial biology, he points to Goethe’s analogies between
living creatures and works of art, and Klee’s understanding of his “picture plane as a kind of
contained, artificial world” [19].
Whitelaw also points to Kasimir Malevich’s Suprematism, which, although it renounced realistic
and imitative painting, viewed the pictorial image as an object in its own right: “a work of pure,
living art,” as Malevich put it. Malevich also considered a machine to be a “technical organism”
and used similar biological metaphors for technology [20].
An Evolutionary Aesthetic?
Do new aesthetic forms in digital art depend on an understanding of the software and
programming? Latham, of course, worked extensively with programmers, but he arrived at the
computer with a strong sense of process in art. His evolutionary sketches show a means of
deploying form in an evolutionary methodology. Because he had been interested in evolving
forms even before he used computers, he was able to apply the most distinctive computer
quality of all: the modelling of dynamic processes.
These artistic systems are not wholly deterministic, running an image through pre-set
parameters until it reaches perfection. Indeed, Latham realized early on that the most interesting
outcomes of his program were quite unforeseen by him: his evolutionary program could arrive
at unexpected conclusions. Even if an artist programs the computer from the start, there will
always be an important element of mystery in the working of the software.
Such quirks render the computer less mechanistic (and predictable) and more “artistic,” because
the outcome of certain operations cannot always be foreseen. This unpredictability can be
harnessed in the same way as the chemical reactions of pigments, or the densities of stone. In
other words, an artist develops a feel for its working and gradually incorporates its idiosyncrasies
into their work, which itself changes subtly or overtly to accommodate these properties.
This is evident in FormSynth and Mutator, where Latham’s choice of operations performed
on the initial shapes guided their eventual appearance. Latham’s stylistic involvement was, in
a sense, pre-visual; it affected the starting point and development of all images generated
through the program rather than just a single artwork. Although he modified of the program’s
underlying code, there were visual consequences because in this way Latham determined the
visual environment in which his shapes could develop. Latham compared the artist to a gardener
who guides the growth of a plant but is not the source of its life. This is itself a new development
for art [21].
372
Lambert, Latham, Fol Leymarie | The Emergence and Growth of Evolutionary Art – 1980-1993
Mitchell Whitelaw sees in this a
factor that is identified in a more
general sense as the artist’s
“signature” or style, because the
“formal vocabulary of elements
and transformations” that takes
place in Latham’s work gives it a
distinctive visual form. However
from a purely exploratory, even
scientific, perspective in terms of
A-Life, it is also a limitation.
The selection Latham exercises is
primarily an aesthetic one that
gives a “non-natural” aspect to
the genetics of his work [22].
Latham’s Organic Art images are
the product of evolutionary
processes and thus indirect
products of his artistic vision.
They are “indirect” in the sense
that Latham developed the
program to evolve shapes along
particular visual lines, but its
continued operation is not
dependent on his intervention.
Like Harold Cohen’s AARON
Figure 5. Timeline 1980 – 2010. © 2012 William Latham and Frederic Fol Leymarie.
simulated painting program [23],
the widely distributed Organic Art software could continue to create Lathamesque images long
after his demise, with varying inputs and changes from computer users. The encoding of his
evolutionary process in software allowed him to make it portable and then distribute it widely as
digital code. Again, this widely distributed software may produce images not directly conceived
by the artist, but the images will be inherent within the parameters of the software. Latham is
responsible for assembling these elements according to his vision and requirements, but the final
image is the result of the software’s own working through these possibilities.
Unlike AARON, however, with its complex relation to Cohen’s creative input, Latham’s software
has a straightforward input procedure and generates images from his initial input parameters.
AARON is not so straightforwardly instructed; it seemingly derives its own decisions about what
to draw from its understanding of art.
There are two different forces at work here. Firstly, there is the artist’s control exercised by
writing or mastering the appropriate software to create images. Secondly, there is the
serendipitous aspect of accidental discovery inherent in an open-ended program where absolute
control yields to experimentation and chance discoveries. In Latham’s work, the evolutionary
nature is the result of a programmer’s control in setting up the initial conditions, then exercising
further choice over the outcomes of these experiments. A fascination with growth and artificial,
yet naturalistic, forms is essential to his art.
The Emergence and Growth of Evolutionary Art – 1980-1993 | Lambert, Latham, Fol Leymarie
373
Conclusion
The period from 1980 to 1993 was initially dominated by fractal art led by Benoit Mandelbrot,
followed by radical developments in evolutionary art as the impact of mutation systems such as
Dawkins’ Biomorphs gained momentum with a number of key evolutionary art figures such as
Karl Sims and the Latham-Todd tandem, whose work was shown extensively at SIGGRAPH
during the period from the late 1980s to the early 1990s. These evolutionary artists were able to
define core rules for growth of computer graphic forms and then use a mutation system that
enabled them to pick and breed purely based on aesthetics to rapidly explore vast areas of
multidimensional parameter space, not knowing what they would discover. Importantly this
work enabled highly intuitive and artist-friendly interfaces to be developed with the addition
of cross-breeding to zoom in on pleasing emerging outcomes.
In parallel to this emergence of evolutionary art using computers, related work was being done
through the 1980s in other scientific and technological areas. In the area of perception and
computer vision, Michael Leyton was, for example, exploring the representation of shapes as
processes that transform them, generating a plausible causal explanation for their “history” [24],
a concept very much akin to Latham’s own thinking in his development of the FormSynth
system. In the area of the computer simulation of plant development, the use of procedural
graphics combined with rule-based L-systems was becoming mature, building from the early
works of biologist Aristid Lindenmayer [25]. Also in parallel, rapid developments were occurring
in evolutionary biology and genetics, in particular from 1990 with the launch of the Human
Genome Project, while in more traditional visual art, strong influences of biology were also
noticeable, for example in the works of Chadwick’s viral landscapes, Hirst’s animal dissections,
and Borland’s preoccupation with anatomy (Figure 5).
Though by the mid-1990s Sims and Latham had become involved in new commercially
focussed projects outside computer art, the core themes and ideas were taken up by new groups
of artists including Jon McCormack and Steven Rooke, and then extended further with the
full emergence of the artificial life research field and associated conferences [26].
Becoming interested again in research, Latham moved back into academia in 2005 and
became professor of computer art at Goldsmiths, University of London in 2007, where he
collaborated with professor Frederic Fol Leymarie and worked again with Stephen Todd after
a gap of 12 years. His recent work has included re-applying the old FormGrow and Mutator
systems rewritten in Java and OpenGL to the world of genomics and scientific visualization
in collaboration with the structural bioinformatics team at Imperial College London [27, 28].
374
Lambert, Latham, Fol Leymarie | The Emergence and Growth of Evolutionary Art – 1980-1993
References
1. Mandelbrot, Benoit, “Fractal Aspects of the Iteration of z —>Lz(L-z) for Complex L and z,” Annals of
the New York Academy of Science, Vol. 357, 249–259 (1980).
2. Mandelbrot, Benoit, The Fractal Geometry of Nature (New York: W.H. Freeman, 1982).
3. Mandelbrot, Benoit, “Fractals and an Art for the Sake of Science,” Leonardo Supplemental Issue, Vol. 2,
Computer Art in Context: SIGGRAPH '89 Art Show Catalog, 21–24 (1989).
4. Turing, Alan M., “The Chemical Basis of Morphogenesis,” Philosophical Transactions of the Royal
Society B, Vol. 237, No. 641, 37–72 (1952).
5. Gardner, Martin, “Mathematical Games: The Fantastic Combinations of John Conway's New Solitaire
Game 'Life',” Scientific American, Vol. 223, 120–123 (1970).
6. Dawkins, Richard, The Blind Watchmaker (New York: Norton & Co., 1986) 55.
7. Levy, Silvano, Desmond Morris: 50 Years of Surrealism (London: Barrie & Jenkins, 1997).
8. Dawkins, Richard, The Blind Watchmaker (New York: Norton & Co., 1986) 8.
9. Ibid., 329.
10. Norton, Alan, “Generation and Display of Geometric Fractals in 3-D,” Computer Graphics, Vol. 16,
No. 3 (1970).
11. Lansdown, John, “The Possible Worlds of William Latham,” The Conquest of Form: Computer Art by
William Latham, Arnolfini Gallery, Bristol, 3 December 1988–15 January 1989.
12. Ibid.
13. Todd, Stephen, and William Latham, Evolutionary Art and Computers (London: Academic Press, 1992) 2.
14. Ibid.
15. Giger, H.R., Giger's Alien (London: Big O Publishing, 1979).
16. Thompson, D'Arcy Wentworth, On Growth and Form (Cambridge: Cambridge Univ Press, 1917).
17. Greenfield, Gary R., “Simulated Aesthetics and Evolving Artworks: A Coevolutionary Approach,”
Leonardo, Vol. 35, No. 3, 283–289 (2002).
18. Whitelaw, Mitchell, “The Abstract Organism: Towards a Prehistory for A-Life Art,” Leonardo, Vol. 34,
No. 4, 345–348 (2001).
19. Ibid., 346.
20. Ibid., 347.
21. Todd, Stephen, and William Latham, Evolutionary Art and Computers (London: Academic Press,
1992) 12.
22. Whitelaw, Mitchell, “Tom Ray's Hammer: Emergence and Excess in A-Life Art,” Leonardo, Vol. 31,
No. 5, 377–381 (1998).
23. McCorduck, Pamela, Meta-Art, Artificial Intelligence, and the Work of Harold Cohen (New York: W.H.
Freeman, 1991).
24. Leyton, Michael, “A Process-Grammar for Shape,” Artificial Intelligence, Vol. 34, No. 2, 213–247 (1988).
25. Prusinkiewicz, Przemyslaw, and Lindenmayer, Aristid, The Algorithmic Beauty of Plants (New York:
Springer-Verlag, 1990).
26. Whitelaw, Mitchell, Metacreation: Art and Artificial Life (Cambridge, MA: MIT Press, 2004).
27. Latham, William, et al., “Using DNA to Generate 3D Organic Art Forms,” Evo'08 Proceedings of the
2008 Conference on Applications of Evolutionary Computing (Berlin: Springer-Verlag, 2008) 433–442.
28. Latham, William, et al., “From DNA to 3D Organic Art Forms,” Proceedings SIGGRAPH '07 ACM
SIGGRAPH 2007 Sketches (New York: ACM, 2007), accessed at <www.siggraph.org/s2007/attendees/
sketches/3.html>.
The Emergence and Growth of Evolutionary Art – 1980-1993 | Lambert, Latham, Fol Leymarie
375